Algorithmic pricing, collusion and the law in 2026

Summary

From March 2024 to November 2026, U.S. enforcers, courts and legislatures set out when shared pricing software turns independent pricing into collusion [2, 7, 8, 10, 11, 12, 18], and European and UK authorities did so from 2019 to 2026 [23, 24, 25, 26, 27, 28].

• U.S. antitrust law reaches pricing software through Section 1 of the Sherman Act, which requires an agreement [1, 10]. On 28 March 2024, the Federal Trade Commission and the Department of Justice said an agreement to use shared pricing recommendations or algorithms is unlawful even where users keep some control over price [2].
• The judgment settling the United States' claims against RealPage, a maker of revenue management software for rental housing, was entered on 19 May 2026. As proposed in November 2025, it bars rivals' nonpublic data while the software sets prices, allows such data in model training only once it is at least 12 months old and not from active leases, and appoints a monitor for three years [4, 7].
• Two federal appeals courts reached different results on shared hotel pricing software. The Ninth Circuit affirmed dismissal in August 2025 where no pooling of confidential data was alleged, and the Third Circuit revived claims in July 2026 where pooling of nonpublic data was alleged [8, 10].
• California, Connecticut, New York and New Jersey have statutes aimed at pricing software, three of them for rental housing [11, 15, 16, 18]. On 29 September 2026, a federal court preliminarily enjoined the section of New York's law aimed at landlords and left standing the section on those who facilitate agreements [11, 12].
• The European Commission's 2023 guidelines on horizontal agreements say firms engaged in illegal pricing cannot avoid liability because algorithms set their prices, and that a pricing rule in a shared tool is likely caught without an explicit agreement [23].
• In experiments, learning pricing agents reached prices above the competitive level without communicating, from Q-learning algorithms in 2020 to LLM pricing agents in 2024 to 2026. A March 2026 preprint found that more competing agents and different algorithms broke such collusion up [29, 30, 31].
• A pre-deployment simulation can show whether pricing agents raise prices above a known competitive level and how prompts and rivals change that [29, 30, 31]. It cannot show legal compliance, because the rulings reviewed turned on what data was pooled, what users agreed to and how often they followed recommendations [7, 8, 10].

Background

Section 1 and the agreement requirement

Section 1 of the Sherman Act, at 15 U.S.C. § 1, declares illegal "Every contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States, or with foreign nations". It sets maximum fines of $100,000,000 for a corporation and $1,000,000 for any other person, and prison of up to 10 years [1]. In the Third Circuit's July 2026 reading, those words require "some form of agreement or 'concerted action'", shown by "unity of purpose", "a meeting of minds" or "conscious commitment to a common scheme" [10]. On circumstantial evidence, parallel conduct needs plus factors: a motive to enter a price-fixing conspiracy, action contrary to the defendant's interests and evidence implying a traditional conspiracy [10].

Shared pricing software can be pleaded as a hub-and-spoke conspiracy, as in the Atlantic City case [10]. The hub is "usually an agent, supplier, or purchaser with whom all the competitors, the spokes, have a relationship", and a per se claim requires "that the spokes are connected to each other via a horizontal agreement", the rim [10]. The Ninth Circuit calls it "simply a collection of vertical and horizontal agreements" [8].

Three ways algorithms can align prices

In November 2019, Germany's Bundeskartellamt and France's Autorité de la concurrence described three scenarios [24]. Algorithms can support a traditional anticompetitive practice that began with human contact. A third party can provide "the same algorithm or somehow coordinated algorithms to competitors", where a central question is whether the competitors knew of its anticompetitive acts "or could at least reasonably have foreseen them". Or collusion can be "induced by the parallel use of individual algorithms" [24]. The UK Competition and Markets Authority (CMA) used similar categories in January 2021, the last being "autonomous tacit collusion, whereby pricing algorithms learn to collude without requiring other information sharing or existing coordination" [26].

The third scenario is the one the agreement requirement may not reach. Article 101 of the Treaty on the Functioning of the European Union prohibits only agreements and concerted practices, so a violation "necessitates some kind of communication", and an algorithm that only observes and reacts to rivals' public behavior "might have to be categorised as intelligent adaptations to the market rather than coordination" [24]. The CMA wrote in 2021 that it was "as yet unclear" whether authorities could object to hub-and-spoke and autonomous tacit collusion without direct contact or a meeting of minds [26]. In March 2026 it added that market investigations may let it remedy coordination without establishing whether a common understanding was reached [27].

This report is for general counsel, compliance leads and pricing teams who use or build pricing software or LLM pricing agents.

Method

Scope. U.S. federal enforcement and courts, four state statutes, one federal bill, the European Commission's guidelines, the French, German and UK authorities, and three experiments. The U.S. matters run from 28 March 2024 to 3 November 2026, the last docket entry seen [14].

Sources. Statutes, bills and status records [1, 11, 16, 17, 18, 20, 21], court opinions, orders and dockets [6, 7, 8, 9, 10, 12, 13, 14], agency releases and filings [2, 3, 4, 5, 15, 19, 22, 23, 24, 25, 26, 27, 28, 33] and papers [29, 30, 31]. A law firm's monthly tracker [32] was used only to find documents.

How. Sources [1] to [32] were opened on 8 November 2026 and [33] on 9 November, and each case status was read from its docket or record on 8 November. Rulings on motions to dismiss take a complaint's allegations as pleaded, so allegations are reported as allegations. Dates computed from a statute's own terms are labelled. The New York opinion is a scanned filing, read by optical character recognition and checked against the docket text. Defendants and software named in public court records appear as subjects of those records.

Findings

1. U.S. enforcers say algorithms are no loophole

On 28 March 2024, the Federal Trade Commission (FTC) and the Department of Justice (DOJ) filed a statement of interest in Cornish-Adebiyi v. Caesars Entertainment in the District of New Jersey [2]. Hotels "cannot use an algorithm to engage in practices that would be illegal if done by a real person", the FTC's release says, and "plaintiffs do not need to identify direct communications between competitors to allege an agreement" [2]. It adds that "an agreement to use shared pricing recommendations, list prices or pricing algorithms is still unlawful" even where participants keep some pricing control, and "Setting or recommending initial starting prices can still violate the antitrust laws" [2]. The Commission voted 3-0-1 [2].

Antitrust Division attorneys appeared for the United States as amicus curiae in Gibson v. Cendyn [8]. Announcing the proposed RealPage settlement on 24 November 2025, Assistant Attorney General Abigail Slater said that "Competing companies must make independent pricing decisions" [3]. On 23 February 2026, the FTC and the DOJ asked for comment on new guidance on business collaborations, naming "algorithmic pricing" among the examples, with a deadline of 24 April 2026 [22], later extended to 21 May 2026 [33]. Their 2000 guidelines had been withdrawn in December 2024 [22].

2. RealPage: the complaint, the consent decrees and the states' claims

The case. On 23 August 2024, the United States and eight states challenged an alleged "unlawful information-sharing scheme" and RealPage's alleged "illegal monopoly in commercial revenue management software" [6]. By 30 September 2026 the plaintiffs were the United States and nine states, Washington having left the case, and the caption named RealPage and six landlords [7]. The amended complaint of 7 January 2025 alleges that RealPage and the landlords agreed to share and use competitively sensitive information and to use RealPage's software to align pricing, in violation of Section 1, and that RealPage monopolized or attempted to monopolize its software market, in violation of Section 2 [4].

Alleged features. The amended complaint alleges that "By enabling auto-accept, a landlord functionally delegates pricing authority to" RealPage, and that RealPage counts a price as accepted when it lands "within 1% of the recommended floor plan price" [4]. The court's 30 September 2026 opinion summarizes the allegations that "nearly 60% of final floor plan prices are within 2.5% of RealPage's recommendation, and more than 85% are within 5%", and that overrides "require written justification and can be escalated to a supervisor" [7].

The judgment. The DOJ announced the proposed final judgment on 24 November 2025, for a 60-day public comment period under the Tunney Act [3], and the court entered it on 19 May 2026 [7]. The proposed judgment's text, in Table 1, reads as a specification for revenue management software [4].

Table 1. The RealPage Proposed Final Judgment as published in the Federal Register on 5 December 2025 [4].

Landlord decrees. As of 30 September 2026, the United States and the plaintiff states had final judgments with Cortland, Greystar and LivCor, and the United States alone had consent decrees with Pinnacle and Willow Bridge, which the court had not adopted [7]. The Willow Bridge decree, published in the Federal Register on 16 July 2026, runs five years from entry and may be terminated after three [5]. If Willow Bridge uses a product from a vendor other than RealPage, it "must secure and submit to the United States a certification from the vendor" that the product meets the decree's Paragraph V.A or the RealPage Final Judgment's requirements [5].

The states' claims. On 30 September 2026, Judge Osteen denied the motions to dismiss of RealPage, Camden, Pinnacle and Willow Bridge, and denied Greystar's and LivCor's as moot [7]. The federal judgment did not moot the states' claims: it can be terminated after four years, does not reach all of the challenged conduct and "does not run to the Plaintiff States", because "The United States alone may enforce the Final Judgment" [7]. On the argument that users were free to override, the court wrote that "the complaint alleges more than simply optional suggestions" [7].

3. Two appeals courts on shared hotel pricing software

Ninth Circuit, 15 August 2025. In Gibson v. Cendyn, hotels had "independently purchased licenses for the same software", which gave pricing recommendations and "did not share any licensing hotel's confidential information among the competing licensees" [8]. The court held that "the choice of several competitors to contract with the same service-provider, when followed by higher prices" is not enough, because "Section 1 requires a causal link between the contested agreement and an anticompetitive restraint of trade in the relevant market", and affirmed dismissal with prejudice [8]. The plaintiffs had alleged a marketing claim that recommendations were "accepted 90% of the time" [8].

Two footnotes mark the holding's limits. The analysis "might change if Plaintiffs had alleged that Cendyn shared the confidential information of each competing hotel among the licensees" [8]. Hotels' freedom to reject recommendations would not immunize them from liability for horizontal price-fixing, but such an agreement would be a hub-and-spoke conspiracy, "any claim of which was waived here" [8]. The Supreme Court denied certiorari on 20 April 2026 [9].

Third Circuit, 29 July 2026. In Cornish-Adebiyi v. Caesars Entertainment, a precedential opinion by Judge McKee reversed the District of New Jersey's dismissal and remanded [10]. The complaint alleges that Atlantic City casino-hotels send "current, non-public room pricing and occupancy data" to Cendyn's Rainmaker software, which processes it with competitors' data into suggested rates "automatically uploaded" to each hotel's platform, and that clients charge the recommended rate "90 percent of the time" [10]. The court held that "the well-pleaded allegations in the CAC are sufficient to support a finding that casino-hotel Defendants have conspired to fix prices through Cendyn's software" [10].

It credited plus factors: hotels followed recommendations "even when it would have been in the casino-hotels' economic interests to reduce room rates", and a former Cendyn vice president urged hotels to "avoid the infamous 'race to the bottom'" [10]. Prices "are fixed when they are agreed upon", whether or not conspirators always adhere to them, and "The alleged 90 percent compliance rate certainly supports an inference of an agreement" [10]. The court made no "assumptions or conclusions about how Cendyn's dynamic pricing software works", and plaintiffs face a higher burden at the next stage [10].

How the two fit. Gibson's plaintiffs did not allege pooling of confidential data and waived the hub-and-spoke count [8]. Cornish-Adebiyi's plaintiffs alleged nonpublic data "consolidated in the hub and then shared back to the spokes in the form of price recommendations" [10]. The Third Circuit cites Gibson only in footnotes 69 and 104, both times for the Nevada district court ruling the District of New Jersey had relied on, and footnote 69 notes the Ninth Circuit's affirmance and the certiorari denial as later history [10].

4. State laws on common pricing algorithms, the New York injunction and a federal bill

New York. Senate Bill S7882 was signed as Chapter 437 on 16 October 2025, adding § 340-b to the General Business Law [11]. A "coordinating function" combines three steps: collecting prices, supply levels or lease dates from two or more landlords, at least two not commonly owned, analyzing them with "a system, software, or process that uses computation, including by using that information to train an algorithm", and recommending rents or other lease terms [11]. Section 2 makes it unlawful "to knowingly or with reckless disregard facilitate an agreement" among landlords not to compete, including by operating or licensing such software. Section 3 makes it unlawful for a landlord to knowingly or recklessly set rents or lease terms "based on recommendations from a software, data analytics service, or algorithmic device performing a coordinating function" [11]. The act took effect on the sixtieth day after it became law, 15 December 2025 by this report's count [11].

The New York injunction. On 29 September 2026, in RealPage v. James, the Southern District of New York granted a preliminary injunction "as to Section 3" and denied the Attorney General's motion to dismiss [12, 14]. The court called it "a close call": the speech is commercial, so a restraint need only survive intermediate scrutiny, "but Section 3 of this Statute, as currently drafted, probably does not" [12]. Section 2 "clarifies that facilitating an agreement between horizontal competitors not to compete is illegal even when the facilitation is done through the use of software", so it survives [12]. The speech Section 3 purportedly restricts, RealPage's "recommendations formulated using almost exclusively public data", "is not per se illegal, so it is protected speech", the court held [12]. The Attorney General moved for reconsideration on 13 October 2026 [13]. The docket shows an opposition on 27 October and a reply on 3 November 2026, and no ruling as viewed on 8 November 2026 [14].

California. Assembly Bill 325 amended the Cartwright Act from 1 January 2026, as the California Energy Commission's Division of Petroleum Market Oversight told gasoline market participants on 21 November 2025 [15]. A common pricing algorithm is "any methodology, including a computer, software, or other technology, used by two or more persons, that uses competitor data to recommend, align, stabilize, set, or otherwise influence a price or commercial term" [15]. Using or distributing one as part of a conspiracy to restrain trade is unlawful, and so is using or distributing one while coercing another person to adopt its recommended price or term [15]. A complaint "shall not be required to allege facts tending to exclude the possibility of independent action" [15].

Connecticut. Section 32 of Public Act 25-1, signed on 26 November 2025, took effect on 1 January 2026 [16, 17]. A "revenue management device" is software that "uses one or more programmed or automated processes to perform calculations of nonpublic competitor data concerning local or state-wide rents or occupancy levels" to advise a landlord on vacancy or rent, and competitor data counts as nonpublic whether attributable to a competitor or anonymized [16]. Using such a device "to set rental rates or occupancy levels for residential dwelling units" is an unlawful practice [16].

New Jersey. The Forbidding the Algorithmic Inflation of Rent (FAIR) Act, P.L. 2026, c.43, was approved on 20 July 2026 [18]. Its coordinating function covers collecting two or more owners' competitively sensitive information to analyze through an algorithm, including by training one, that sets or recommends prices, lease terms or occupancy, setting prices through an algorithm that processes another owner's sensitive information, and setting or recommending terms for two or more owners through the same or a substantially similar algorithm "that facilitates parallel pricing coordination" [18]. Parallel pricing coordination is "a tacit or express agreement" to raise, change, maintain or manipulate pricing, and a tacit agreement is "mutual consent demonstrated without verbal or written communication" [18]. Section 4 makes unlawful, under the New Jersey Antitrust Act, paying for a coordinator's services, facilitating such an agreement, parallel pricing coordination and performing a coordinating function [18]. Under Section 9 the act takes effect on the first day of the twelfth month after enactment, 1 July 2027 by this report's count [18]. The Governor's release calls New Jersey "just the fourth state to explicitly regulate rent setting algorithms" [19].

A federal bill. S. 232, the Preventing Algorithmic Collusion Act of 2025, was introduced on 23 January 2025 by Senator Klobuchar with eight cosponsors and referred to the Committee on the Judiciary, the last action in its status record [20, 21]. Section 4(a) would make it unlawful "to use or distribute any pricing algorithm that uses, incorporates, or was trained with nonpublic competitor data", and Section 5(a) would presume a Section 1 agreement where the plaintiff shows, among other routes, that the defendant distributed such an algorithm to two or more persons intending it to set or recommend prices in the same or a related market [20].

5. European and UK views, including agents

The European Commission. The 2023 Horizontal Guidelines say firms "can for instance independently use algorithms to monitor the prices of competitors and to inform their own price setting", while "Collusion by code on essential parameters of competition is typically a cartel and therefore a restriction of competition by object" [23]. Firms in illegal pricing practices "cannot avoid liability on the ground that their prices were determined by algorithms", and like an employee or outside consultant, "an algorithm remains under the firm's control" [23]. A pricing rule in a shared algorithmic tool, such as matching a particular competitor's price minus 5%, "is also likely to be caught by Article 101(1), even in the absence of an explicit agreement to align future pricing" [23]. Aggregating commercially sensitive information in a pricing tool that one IT company offers to several competitors "could amount to horizontal collusion", and a firm sharing commercially sensitive information through a third party may be infringing Article 101 where it "could reasonably have foreseen" onward sharing "and it was prepared to accept the risk" [23].

The UK CMA. On 4 March 2026, the CMA wrote that "UK law recognises that indirect information exchange through third parties, such as algorithm providers, can constitute illegal, anti-competitive conduct" [27]. It recalled a 2016 case in which two online poster sellers agreed not to undercut each other and gave effect to that agreement with pricing software [27]. Fines can reach 10% of annual turnover, leniency covers exchanging sensitive information through a shared algorithm, informants can receive up to £250,000, and the CMA is "actively screening for algorithmic collusion" [27]. For pricing solutions involving LLMs, it suggests "linguistic stress testing prompts" and "explicit anti-collusion constraints" [27].

The CMA's paper "Agentic AI and consumers", first published on 9 March 2026 and updated on 30 October 2026, says that where several businesses deploy autonomous pricing agents, "interaction between these systems could dampen competitive pressure", and that businesses "must take proactive steps to understand, test and govern the technologies they deploy" [28]. In 2021, the CMA had called the real-world risk unclear "due to a relative paucity of empirical evidence", and reported German retail petrol stations that "increased their margins by around 9 percent after adopting algorithmic pricing, but only where they faced local competition" [26].

France. The Autorité de la concurrence issued Opinion 26-A-05 on 17 July 2026, after about 40 contributions to its consultation [25]. Stakeholders raised algorithmic collusion, a risk the authority called "not new" that could materialise in agentic commerce "if AI agents ... were able, in particular, to participate in price negotiations" [25]. The first of its six recommendations is that "The existing regulatory framework should be implemented fully and effectively", with soft-law instruments first and any regulation at European Union level [25].

6. What experiments show about learning pricing agents

Calvano, Calzolari, Denicolò and Pastorello (American Economic Review, October 2020) found that Q-learning algorithms in a standard model of repeated price competition consistently learned to charge supracompetitive prices without communicating, sustained by finite punishment followed by a gradual return to cooperation, across cost and demand asymmetries and changes in the number of players [29].

Fish, Gonczarowski and Shorrer (first posted in March 2024, version 6 of 31 August 2026, accepted to EC 2026) found that LLM-based pricing agents in oligopoly settings reach supracompetitive prices and profits quickly and autonomously, and that small variations in prompt wording substantially change the degree of supracompetitive pricing. The results extend to auctions [30].

Keppo, Li, Tsoukalas and Yuan (preprint, 18 March 2026) ran open-source LLM agents for over 2,000 compute hours [31]. Differences in patience cut the price lift from 22% to 10% above competitive levels, and asymmetric data access cut it to 7% [31]. More competing agents broke up collusion, and so did different algorithms, while differences in model size stabilized it [31]. The CMA cites this work and calls the evidence early, with no overall consensus [27]. The review of LLM agents as market participants covers the Calvano and Fish studies and related work.

The agreement tests reviewed turn on pooled nonpublic data, an intermediary or communication [8, 10, 23, 24]. Among the texts reviewed, New Jersey's act is the one whose definitions reach tacit agreement, and it applies to rental housing [18].

7. What a pre-deployment simulation can and cannot show

What the sources ask of firms. The CMA asks businesses to "understand, test and govern" what they deploy [28]. The joint study concluded that firms "need to think about how they could ensure antitrust compliance when using pricing algorithms" [24].

What a simulation can show. The experiments measure three things a simulated market can repeat for a given agent: whether prices rise above a known competitive benchmark without instruction [29, 30], how much instruction wording moves the result [30], and how the number and mix of rival agents change it [31]. In a simulated market the competitive equilibrium is known before the first bid, so the distance between it and the prices agents set can be measured.

What it cannot show. A run does not show legal compliance. The rulings reviewed turned on what data was pooled, what users agreed to and how often they followed recommendations [7, 8, 10], and a laboratory run establishes none of those facts. The joint study left open whether pricing algorithms could align "by chance" under real market conditions [24]. A simulation result is behavioral evidence for counsel to weigh.

Implications

For general counsel and compliance leads. The texts reviewed ask about the same facts: whether rivals' data reaches the software in operation or in training, nonpublic data in most of them and any two landlords' data in New York's [4, 10, 11, 16, 18, 23], whether users agreed to use its prices [2, 10], how often they followed them and what an override costs [4, 7, 10], and whether a firm could foresee that a vendor would pool its data [23, 24]. Three matters had no final outcome in the records reviewed: the New York reconsideration motion, with no ruling on the docket on 8 November 2026 [14], the states' RealPage claims, which survived dismissal on 30 September 2026 [7], and the Atlantic City case, remanded on 29 July 2026 [10].

For pricing teams and software vendors. Table 1 lists the features a court decree required one revenue management product to change [4]. Through the Willow Bridge decree, those terms are also a standard other vendors certify against for one landlord [5]. California's definition of a common pricing algorithm names no industry [15].

For teams deploying LLM pricing agents. The CMA names LLM pricing directly [27]. The experiments show prompt wording and the mix of rivals moving prices [30, 31], so a test with one prompt and one rival measures one case.

Pre-deployment testing. Rillor's market simulation measures whether agents coordinate, tacitly or openly, to move prices, and whether what they tell other agents matches what they do. Every measure is computed from the run log: bids, messages, prices and outcomes.

Limits of this analysis

• This report is research on what statutes, court records and agency statements say. It is not legal advice and draws no legal conclusion about any product or practice.
• Statuses are as viewed on 8 November 2026, and the three matters listed under Implications can change them [7, 10, 14].
• The rulings reviewed decide motions to dismiss or a preliminary injunction. Quoted allegations are allegations, and no ruling reviewed finds that any defendant fixed prices.
• The scope is U.S. federal law, four state statutes, the EU guidelines and the French, German and UK authorities. Other states' laws, private suits beyond those named and laws on individualized pricing are outside it.
• The French opinion is described from the Autorité's press release [25], the petrol-station study from the CMA's account [26], and California's AB 325 from the Energy Commission's letter, which quotes the statute [15].
• This report covers the collaboration guidance only to its comment deadline, 21 May 2026 [33].
• The experiments are laboratory studies in stylized markets, two of them preprints [30, 31].

Sources

1. 15 U.S. Code § 1, Trusts, etc., in restraint of trade illegal, penalty. Legal Information Institute, Cornell Law School, viewed 8 November 2026. https://www.law.cornell.edu/uscode/text/15/1
2. FTC and DOJ File Statement of Interest in Hotel Room Algorithmic Price-Fixing Case, press release. U.S. Federal Trade Commission, 28 March 2024, viewed 8 November 2026. https://www.ftc.gov/news-events/news/press-releases/2024/03/ftc-doj-file-statement-interest-hotel-room-algorithmic-price-fixing-case
3. Justice Department Requires RealPage to End the Sharing of Competitively Sensitive Information and Alignment of Pricing Among Competitors, press release 25-1111. U.S. Department of Justice, Office of Public Affairs, 24 November 2025, viewed 8 November 2026. https://www.justice.gov/opa/pr/justice-department-requires-realpage-end-sharing-competitively-sensitive-information-and
4. United States of America et al. v. RealPage, Inc. et al., Proposed Final Judgment and Competitive Impact Statement (RealPage, Inc.), 90 FR 56286. U.S. Department of Justice, Antitrust Division, Federal Register, 5 December 2025, viewed 8 November 2026. https://www.govinfo.gov/content/pkg/FR-2025-12-05/pdf/2025-21966.pdf
5. United States of America, et al. v. RealPage, Inc., et al., Proposed Final Judgment and Competitive Impact Statement (Willow Bridge Property Company), 91 FR 43774. Federal Register, 16 July 2026, viewed 8 November 2026. https://www.govinfo.gov/content/pkg/FR-2026-07-16/pdf/2026-14345.pdf
6. United States of America et al. v. RealPage Inc., No. 1:24CV710, Memorandum Opinion and Order (motion to transfer venue), Doc. 45. U.S. District Court for the Middle District of North Carolina, 20 December 2024, via GovInfo, viewed 8 November 2026. https://www.govinfo.gov/content/pkg/USCOURTS-ncmd-1_24-cv-00710/pdf/USCOURTS-ncmd-1_24-cv-00710-0.pdf
7. United States of America et al. v. RealPage, Inc. et al., No. 1:24-cv-710, Memorandum Opinion and Order (motions to dismiss), Doc. 223. U.S. District Court for the Middle District of North Carolina, 30 September 2026, via GovInfo, viewed 8 November 2026. https://www.govinfo.gov/content/pkg/USCOURTS-ncmd-1_24-cv-00710/pdf/USCOURTS-ncmd-1_24-cv-00710-1.pdf
8. Gibson v. Cendyn Group, LLC, No. 24-3576, opinion. U.S. Court of Appeals for the Ninth Circuit, 15 August 2025, viewed 8 November 2026. https://cdn.ca9.uscourts.gov/datastore/opinions/2025/08/15/24-3576.pdf
9. Docket No. 25-1109, Richard Gibson, et al., Petitioners v. Cendyn Group, LLC, et al. Supreme Court of the United States, viewed 8 November 2026. https://www.supremecourt.gov/search.aspx?filename=/docket/docketfiles/html/public/25-1109.html
10. Cornish-Adebiyi v. Caesars Entertainment, Inc., No. 24-3006, precedential opinion. U.S. Court of Appeals for the Third Circuit, 29 July 2026, viewed 8 November 2026. https://www2.ca3.uscourts.gov/opinarch/243006p.pdf
11. Senate Bill S7882 (2025-2026), summary, actions and text (General Business Law § 340-b). New York State Assembly, viewed 8 November 2026. https://nyassembly.gov/leg/?default_fld=&leg_video=&bn=S07882&term=2025&Summary=Y&Actions=Y&Text=Y
12. RealPage, Inc. v. James, No. 1:25-cv-09847-VEC, Opinion and Order, Doc. 52. U.S. District Court for the Southern District of New York, 29 September 2026, via CourtListener (RECAP), viewed 8 November 2026. https://storage.courtlistener.com/recap/gov.uscourts.nysd.653712/gov.uscourts.nysd.653712.52.0.pdf
13. RealPage, Inc. v. James, No. 1:25-cv-09847-VEC, Notice of Motion for Reconsideration, Doc. 56. U.S. District Court for the Southern District of New York, 13 October 2026, via CourtListener (RECAP), viewed 8 November 2026. https://storage.courtlistener.com/recap/gov.uscourts.nysd.653712/gov.uscourts.nysd.653712.56.0.pdf
14. RealPage, Inc. v. James, 1:25-cv-09847, docket. CourtListener, Free Law Project, viewed 8 November 2026. https://www.courtlistener.com/docket/71964352/realpage-inc-v-james/
15. Re: Assembly Bill 325, effective January 1, 2026, letter to gasoline market participants. California Energy Commission, Division of Petroleum Market Oversight, 21 November 2025, viewed 8 November 2026. https://www.energy.ca.gov/sites/default/files/2026-05/DPMO_AB325_Notice_ada.pdf
16. November 2025 Special Session, Public Act No. 25-1 (House Bill No. 8002), An Act Concerning Housing Growth, Sec. 32. Connecticut General Assembly, 2025, viewed 8 November 2026. https://www.cga.ct.gov/2025/ACT/PA/PDF/2025PA-00001-R00HB-08002SS1-PA.PDF
17. Bill status, HB 8002 (2025), An Act Concerning Housing Growth. Connecticut General Assembly, viewed 8 November 2026. https://www.cga.ct.gov/asp/cgabillstatus/cgabillstatus.asp?selBillType=Bill&which_year=2025&bill_num=8002
18. P.L. 2026, c.43 (A3497 1R ACS), Forbidding the Algorithmic Inflation of Rent (FAIR) Act. New Jersey Legislature, approved 20 July 2026, viewed 8 November 2026. https://pub.njleg.state.nj.us/Bills/2026/PL26/43_.HTM
19. Governor Sherrill Signs the Forbidding the Algorithmic Inflation of Rent (FAIR) Act Into Law to Protect Renters from Algorithmic Rent Setting. State of New Jersey, Office of the Governor, 20 July 2026, viewed 8 November 2026. https://www.nj.gov/governor/news/2026/20260720a.shtml
20. S. 232 (119th Congress), Preventing Algorithmic Collusion Act of 2025, introduced in Senate. U.S. Government Publishing Office, GovInfo, 23 January 2025, viewed 8 November 2026. https://www.govinfo.gov/content/pkg/BILLS-119s232is/html/BILLS-119s232is.htm
21. Bill status, S. 232 (119th Congress), BILLSTATUS XML. U.S. Government Publishing Office, GovInfo, updated 10 August 2026, viewed 8 November 2026. https://www.govinfo.gov/bulkdata/BILLSTATUS/119/s/BILLSTATUS-119s232.xml
22. Federal Trade Commission and Department of Justice Seek Public Comment for Guidance on Business Collaborations, press release. U.S. Federal Trade Commission, 23 February 2026, viewed 8 November 2026. https://www.ftc.gov/news-events/news/press-releases/2026/02/federal-trade-commission-department-justice-seek-public-comment-guidance-business-collaborations
23. Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreements (2023/C 259/01). European Commission, Official Journal of the European Union, 21 July 2023, viewed 8 November 2026. https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:52023XC0721(01)
24. Algorithms and Competition. Bundeskartellamt and Autorité de la concurrence, November 2019, viewed 8 November 2026. https://www.bundeskartellamt.de/SharedDocs/Publikation/EN/Berichte/Algorithms_and_Competition_Working-Paper.pdf?__blob=publicationFile&v=5
25. AI agents: the Autorité de la concurrence issues its opinion on the competitive functioning of the AI agents sector (Opinion 26-A-05), press release. Autorité de la concurrence, 17 July 2026, viewed 8 November 2026. https://www.autoritedelaconcurrence.fr/en/press-release/ai-agents-autorite-de-la-concurrence-issues-its-opinion-competitive-functioning-ai
26. Algorithms: How they can reduce competition and harm consumers. UK Competition and Markets Authority, 19 January 2021, viewed 8 November 2026. https://www.gov.uk/government/publications/algorithms-how-they-can-reduce-competition-and-harm-consumers/algorithms-how-they-can-reduce-competition-and-harm-consumers
27. AI and collusion: frontiers, opportunities and challenges. Karen Croxson and Juliette Enser, Competition and Markets Authority blog, 4 March 2026, viewed 8 November 2026. https://competitionandmarkets.blog.gov.uk/2026/03/04/ai-and-collusion-frontiers-opportunities-and-challenges/
28. Agentic AI and consumers. UK Competition and Markets Authority, first published 9 March 2026, updated 30 October 2026, viewed 8 November 2026. https://www.gov.uk/government/publications/agentic-ai-and-consumers/agentic-ai-and-consumers
29. Artificial Intelligence, Algorithmic Pricing, and Collusion. Emilio Calvano, Giacomo Calzolari, Vincenzo Denicolò and Sergio Pastorello, American Economic Review 110(10), 3267-97, American Economic Association, October 2020, viewed 8 November 2026. https://www.aeaweb.org/articles?id=10.1257/aer.20190623
30. Algorithmic Collusion by Large Language Models. Sara Fish, Yannai A. Gonczarowski and Ran I. Shorrer, arXiv:2404.00806, version 6, 31 August 2026 (first version 31 March 2024, accepted to EC 2026), viewed 8 November 2026. https://arxiv.org/abs/2404.00806
31. On the Fragility of AI Agent Collusion. Jussi Keppo, Yuze Li, Gerry Tsoukalas and Nuo Yuan, arXiv:2603.20281, version 1, 18 March 2026, viewed 8 November 2026. https://arxiv.org/abs/2603.20281
32. Latest Intelligence: Antitrust Developments Impacting Providers and Users of Algorithmic Tools (September 2026). Fiona A. Schaeffer and others, The Milbank General Counsel Blog, Milbank LLP, 11 October 2026, viewed 8 November 2026. https://www.milbankgeneralcounsel.com/2026/10/latest-intelligence-antitrust-developments-impacting-providers-and-users-of-algorithmic-tools-september-2026/
33. DOJ and FTC Extend Deadline for Public Comment on Guidance on Business Collaborations, press release 26-373. U.S. Department of Justice, Office of Public Affairs, 17 April 2026, viewed 9 November 2026. https://www.justice.gov/opa/pr/doj-and-ftc-extend-deadline-public-comment-guidance-business-collaborations